An Update to the Kaiser Permanente Inpatient Risk Adjustment Methodology Accurately Predicts In-Hospital Mortality: a Retrospective Cohort Study.
in-hospital mortality
inpatient care
risk adjustment
troponin
validation
Journal
Journal of general internal medicine
ISSN: 1525-1497
Titre abrégé: J Gen Intern Med
Pays: United States
ID NLM: 8605834
Informations de publication
Date de publication:
Nov 2023
Nov 2023
Historique:
received:
10
01
2023
accepted:
16
05
2023
pmc-release:
01
11
2024
medline:
28
11
2023
pubmed:
10
6
2023
entrez:
9
6
2023
Statut:
ppublish
Résumé
Methods to accurately predict the risk of in-hospital mortality are important for applications including quality assessment of healthcare institutions and research. To update and validate the Kaiser Permanente inpatient risk adjustment methodology (KP method) to predict in-hospital mortality, using open-source tools to measure comorbidity and diagnosis groups, and removing troponin which is difficult to standardize across modern clinical assays. Retrospective cohort study using electronic health record data from GEMINI. GEMINI is a research collaborative that collects administrative and clinical data from hospital information systems. Adult general medicine inpatients at 28 hospitals in Ontario, Canada, between April 2010 and December 2022. The outcome was in-hospital mortality, modeled by diagnosis group using 56 logistic regressions. We compared models with and without troponin as an input to the laboratory-based acute physiology score. We fit and validated the updated method using internal-external cross-validation at 28 hospitals from April 2015 to December 2022. In 938,103 hospitalizations with 7.2% in-hospital mortality, the updated KP method accurately predicted the risk of mortality. The c-statistic at the median hospital was 0.866 (see Fig. 3) (25th-75th 0.848-0.876, range 0.816-0.927) and calibration was strong for nearly all patients at all hospitals. The 95th percentile absolute difference between predicted and observed probabilities was 0.038 at the median hospital (25th-75th 0.024-0.057, range 0.006-0.118). Model performance was very similar with and without troponin in a subset of 7 hospitals, and performance was similar with and without troponin for patients hospitalized for heart failure and acute myocardial infarction. An update to the KP method accurately predicted in-hospital mortality for general medicine inpatients in 28 hospitals in Ontario, Canada. This updated method can be implemented in a wider range of settings using common open-source tools.
Sections du résumé
BACKGROUND
BACKGROUND
Methods to accurately predict the risk of in-hospital mortality are important for applications including quality assessment of healthcare institutions and research.
OBJECTIVE
OBJECTIVE
To update and validate the Kaiser Permanente inpatient risk adjustment methodology (KP method) to predict in-hospital mortality, using open-source tools to measure comorbidity and diagnosis groups, and removing troponin which is difficult to standardize across modern clinical assays.
DESIGN
METHODS
Retrospective cohort study using electronic health record data from GEMINI. GEMINI is a research collaborative that collects administrative and clinical data from hospital information systems.
PARTICIPANTS
METHODS
Adult general medicine inpatients at 28 hospitals in Ontario, Canada, between April 2010 and December 2022.
MAIN MEASURES
METHODS
The outcome was in-hospital mortality, modeled by diagnosis group using 56 logistic regressions. We compared models with and without troponin as an input to the laboratory-based acute physiology score. We fit and validated the updated method using internal-external cross-validation at 28 hospitals from April 2015 to December 2022.
KEY RESULTS
RESULTS
In 938,103 hospitalizations with 7.2% in-hospital mortality, the updated KP method accurately predicted the risk of mortality. The c-statistic at the median hospital was 0.866 (see Fig. 3) (25th-75th 0.848-0.876, range 0.816-0.927) and calibration was strong for nearly all patients at all hospitals. The 95th percentile absolute difference between predicted and observed probabilities was 0.038 at the median hospital (25th-75th 0.024-0.057, range 0.006-0.118). Model performance was very similar with and without troponin in a subset of 7 hospitals, and performance was similar with and without troponin for patients hospitalized for heart failure and acute myocardial infarction.
CONCLUSIONS
CONCLUSIONS
An update to the KP method accurately predicted in-hospital mortality for general medicine inpatients in 28 hospitals in Ontario, Canada. This updated method can be implemented in a wider range of settings using common open-source tools.
Identifiants
pubmed: 37296357
doi: 10.1007/s11606-023-08245-w
pii: 10.1007/s11606-023-08245-w
pmc: PMC10682304
doi:
Substances chimiques
Troponin
0
Types de publication
Journal Article
Langues
eng
Sous-ensembles de citation
IM
Pagination
3303-3312Informations de copyright
© 2023. The Author(s), under exclusive licence to Society of General Internal Medicine.
Références
BMC Health Serv Res. 2011 Oct 07;11:258
pubmed: 21982489
J Am Coll Cardiol. 2019 Mar 12;73(9):1059-1077
pubmed: 30798981
Health Care Law Mon. 2011 Feb;2011(2):2-9
pubmed: 21400963
Qual Manag Health Care. 2016 Jul-Sep;25(3):123-8
pubmed: 27367212
Clin Chem. 2010 Feb;56(2):254-61
pubmed: 19959623
Stat Med. 1996 Feb 28;15(4):361-87
pubmed: 8668867
Stat Med. 2019 Sep 20;38(21):4051-4065
pubmed: 31270850
CMAJ Open. 2017 Dec 11;5(4):E842-E849
pubmed: 29237706
J Clin Epidemiol. 2001 Aug;54(8):774-81
pubmed: 11470385
J Am Med Inform Assoc. 2021 Mar 1;28(3):578-587
pubmed: 33164061
Crit Care Med. 2007 Apr;35(4):1091-8
pubmed: 17334248
J Gen Intern Med. 2018 Nov;33(11):1899-1904
pubmed: 30054888
BMC Cardiovasc Disord. 2021 Nov 30;21(1):571
pubmed: 34847863
JAMA Netw Open. 2019 Jul 3;2(7):e197314
pubmed: 31314120
Med Care. 2008 Mar;46(3):232-9
pubmed: 18388836
J Clin Epidemiol. 2010 Jul;63(7):798-803
pubmed: 20004550
JAMA Netw Open. 2019 Dec 2;2(12):e1916769
pubmed: 31800072
JAMA Netw Open. 2022 Jun 1;5(6):e2218172
pubmed: 35737389
J Biomed Inform. 2016 Dec;64:10-19
pubmed: 27658885
PLoS One. 2014 Jul 07;9(7):e102046
pubmed: 25000586
Am J Epidemiol. 2011 Mar 15;173(6):676-82
pubmed: 21330339
Med Care. 2010 Aug;48(8):739-44
pubmed: 20613662
BMJ Open. 2015 Jun 05;5(6):e007974
pubmed: 26048212
Med Care. 2013 May;51(5):446-53
pubmed: 23579354
J Anim Ecol. 2015 Jul;84(4):892-7
pubmed: 26074184
J Clin Epidemiol. 2021 Sep;137:83-91
pubmed: 33836256
BMC Med Res Methodol. 2021 Jan 7;21(1):9
pubmed: 33413132
J Gen Intern Med. 2022 Nov;37(15):3877-3884
pubmed: 35028862
J Am Heart Assoc. 2019 Oct;8(19):e013551
pubmed: 31547767
Crit Care Med. 2006 May;34(5):1297-310
pubmed: 16540951
Circ Heart Fail. 2016 Aug;9(8):
pubmed: 27514749